Search code examples
c#jsonserializationjson.netjson-deserialization

Remove illegal character from JSON before deserializing with JSON.NET


I have a string like which I want deserialize to datatable. Because it has a double quote in the value JSON.NET is not able to deserialize this.

[{"name":"soham "the" dasgupta"}]

Now I would like to replace the "the" with the. Please help. Replace("\"", "") is not working obviously.


Solution

  • Basically you need to find the characters between the quotes and check if they are valid JSON delimiters.

    I've wrote a quick and dirty function which searches the substrings between the strings and matches them with valid JSON delimiters. If the substring don't match, it escapes the quotes.

    Note: This code is an very basic example of my idea, it could possibly not deal with every kind of JSON input.

    static string FixJson(string input)
    {
        var output = input;
        for (var x = 0; x < input.Length; x++)
        {
            if (input[x] != '\"') continue;
    
            for (var y = x + 1; y < input.Length; y++)
            {
                if (input[y] != '\"') continue;
    
                var found = false;
                for (var z = y + 1; z < input.Length; z++)
                {
                    if (input[z] != '\"') continue;
    
                    var tmp = input.Substring(y + 1, z - y - 1);
                    if (tmp.Any(t => t != ' ' && t != ':' && t != ',' && t != '{' && t != '}'))
                    {
                        output = output.Replace("\"" + tmp + "\"", "\\\"" + tmp + "\\\"");
                    }
    
                    x = y;
                    found = true;
                    break;
                }
    
                if (found)
                    break;
            }
        }
    
        return output;
    }